How to Create More Capacity in Data Center

With the ever-increasing demand for more computing power and data center servers, data center manager have the responsibility to chart a data center capacity plan and determine what strategy will accommodate business needs best. Of course, they could just expand to large facilities (upgrading to the advanced switch and fiber enclosures). However, not all IT budgets are increasing, and many users just cannot afford the extra money. Therefore, people are turning to high-density and cost-effective infrastructures. To support these applications, this article will introduce a broad selection of high-density connectivity and high capacity cable management devices.

• High-Density Patch Panels

High-Density (HD) fiber patch panel solution is the most convenient approach for solving the problem of limited capacity in a data center environment. And it provides a flexible way to connect devices of different generations of equipment quickly and easily. HD patch panels consist of a panel enclosure and modular HD cassettes, which can connect a fiber network feed (via multi-strand or MTP cable) and segment it into standard LC connections in order to interface with 10Gbps devices. The following image shows a fiber adapter panel with 12 LC duplex single-mode adapters.

HD fiber patch panels feature the following advantages:
Flexibility: They can connect different generations of equipment such as 10Gb, 40Gb, 100Gb in a simple panel-cassette system with different connector types.

Ease of Installation: No tools are required to install the cassette in the panel enclosure. Each cassette features factory terminated connectors that reduce the time and labor required of field connector terminations.

Cost-Effectiveness: High-density and ease of installation provide a low initial investment cost. Flexibility, adjustability and reliability provide a high ROI. What’s more, network reconfiguration is highly adjustable due to the modular cassette system.

• High-Density Patch Cords

As cabling density increases along with the deployment of higher network speeds, HD patch cords deliver a robust design to withstand the rigorous of daily use. Cables that can offer a smaller overall diameter improve cable management by installing in dense patch cord trays that take up less space. They also provide better airflow to maintain consistent operating temperatures, reducing the likelihood of failure or downtime.

Finger access to each patch cable, furthermore hinder the cable management and makes the cable installation become more difficult. To ensure easy access, high-density patch cords are easy to remove through the use of a flexible pull-tab fiber optic cable just as seen in the above picture. This cable type has the same component and internal structure as the traditional patch cords (e.g.SC FC patch cord), except the a tab attached to the connector, which makes it easy for cable management. These tabs can help increase cabling density and maintain reliability, preventing you from accidentally loosening surrounding connectors as you access the patch cord you need.

MPO/MTP trunk cable is the another example of the HD patch cords. These cables are the foundation of easier, faster and better pre-terminated fiber connectivity solutions, as it allows tighter trunk cable bends for slack storage and routing. With the high-density trunks in your data center solution, less space is consumed and installation is easier.

• High-Density Fiber Enclosures

Fiber enclosure makes full used of the spaces in data center by combining most of the fiber optic connections in strong standards modules, providing solid protection of data center links and increasing cabling density. Therefore, data center managers can get easy access to fiber connections and easy cable management. Accordingly, the cost for data center installation and maintenance can be effectively reduced.

Fiber enclosures are usually available in 1U, 2U, 3U, 4U. The 1U rack mount fiber enclosure is the most commonly used one on the market. Now 4U or larger rack mount fiber enclosures are also becoming popular driven by the increasing of fiber counts in data center. Except standards rack mount fiber enclosures, a lot of data centers or server rooms use customized fiber enclosures for their special requirements.

FS.COM FHD Series rack mount fiber enclosures are available in 1RU, 2RU and 4RU rack unit options. With optional FHD cassette modules or adapter panels in single-mode, multimode, or 10G multimode versions, users can install, maintain, and upgrade their cabling systems in a more flexible and cost-effective way. In addition to rack mount solution, our FHD series products also support wall mount type which can meet the cabling demands on fiber industrial environments.

Conclusion

High-density optic solutions enable data center operators to maximize the amount of active equipment and cables in a data center by minimizing the foot print of the networking infrastructure. Besides the above HD optical products, there are also a range of HD products including the high speed interconnect optics, cable assemblies, cable management hardware. If you want to know more about the HD products from FS.COM, please have a look at our website.

10G Data Center Cabling Solutions

The rapid development in data center throughput has led to the increasing usage and demand for higher-performance servers, storage and interconnects. And the old 1G Ethernet cannot handle the heavy-loaded solutions in data center any more. As a result, datacenter designers are looking to the expansion of higher speed Ethernet solutions, specifically 10 G and 40G Ethernet. As for 10GbE, there are two broad categories—SFP+ optical options and 10GBASE-T available on the market, which pose difficulty in selecting the appropriate 10-gigabit physical media. This article will make a brief introduction to these two 10G solutions to help you choose the suitable one.

What Is 10G SFP+?

10G Small-form-factor pluggable plus (SFP+) is the industry standard for data rates up to 10 Gbps, which is also MSA compliant. SFP+ module is especially standardized for 10 Gbps application, and is identical in dimensions to the SFP. To achieve the SFP+ form factor reduction, only the optical-to-electrical and electrical-to-optical conversion functionalities occur inside the optical module. The key advantage of SFP+ over the existing 10 G optical interconnects is the higher port densities enabled by its dimensions, and lower power consumption. Figure 1 shows a SFP+ modules connected with a LC to LC patch cord.

SFP+ transceivers are available in different 10G Ethernet standards—10GBASE-SR, 10GBASE-LR, 10GBASE-LRM, 10GBASE-ER, 10GBASE-ZR, 10GBASE-LX4 and 10GBASE-PR. Each standard has a unique specification that can be suitable for different applications. SFP+ optics are selected more often when designers need faster and more reliable solutions to handle 10 Gigabit Ethernet optical lines. With lower power usage and low latency, SFP+ ports are most commonly used for enterprise switches and also for plug-in cards for servers.

How Does 10G Copper Solutions Compare?

10G Ethernet can also run over twin-axial cabling, twisted pair cabling, and backplanes. SFP+ direct attach cable, 10GbASE-CX4 and 10GBASE-T are the common copper solutions for 10G short-reach interconnect. 10GBase-CX4 achieves the aggregate rate of 10 Gbps by transmitting over four independent cables at 3.125 Gbps. The disadvantage of this solution is the bulkiness of the cables that have eight twin-ax cables within them for a duplex link. And the bulkiness of the cables makes cable management more difficult. This becomes a significant limitation as the port densities and inter-port connection densities increase. Additionally, the number of conductors make these cables expensive compared to SFP+ cables.

10GBase-T enables 10 Gbps transmissions over Cat 6 and higher quality cables using complex signal processing and channel coding. The potential advantage of this technology is its extended reach. The extended reach and ability to enable structured cabling are not required for the short reach interconnects between servers and switches collocated in a rack. The disadvantages are its high power consumption and latency. The high latency in particular is a key limitation in latency sensitive data center and storage applications.

SFP+ Direct Attach DAC is another lower cost alternative to fiber with a limited link length of 7 meters. Additionally, it has significant power, cost, and performance advantages over the above media as explained below.

Why Use 10G DAC Cables in Data Center?

As 10 Gbps interconnects become ubiquitous in servers, providers are looking for a low cost, low power, and space efficient interconnect solution for the short reach (5-7m) links that dominate the data center environment. A length of seven meters covers all connections between server cards and switches, typically mounted on a single data center rack, and a vast majority of inter-rack connections. The SFP+ Direct Attach Copper is the 10G interconnect technology that matches these requirements, playing a vital role in enabling the next generation power and cost efficient data centers.

SFP+ Direct Attach Copper Solution

To further reduce cost and power in interconnect distances of 7 meters, which is sufficient to link server cards and switches, the SFP+ Direct Attach Copper replaces the optical modules and fiber with a passive copper cable with connectors identical to an SFP+ optical module. Figure 2 shows a SFP-H10GB-CU3M plugging in a Cisco switch. The reduction in cost and power are significant since the price of two optical modules required to support a full duplex link is approximately 10 times the cost of a fully connectorized SFP+ Direct Attach cable. In addition, the optical SFP+ modules consume around 1W each, adding 2W per port to the overall system power budget and cooling requirements.

Data is transmitted as 10 G serial NRZ (Non Return to Zero) symbols with transmit pre-emphasis and receive equalization compensating for Inter Symbol Interference caused by the board trace and the copper cable. This choice of serial NRZ transmission over a passive medium makes SFP+ copper both a low power and low latency solution compared to its alternatives.

Conclusion

As more emphasis is placed on energy efficient data centers and higher bandwidth applications, the need for a small form factor, low power, low latency and low cost interconnect makes SFP+ Direct Attach the optimal solution for short-reach 10G interconnects. While SFP+ fiber options provides a great path for higher performance long haul applications. FS.COM is a top manufacturer for DAC Twinax cables (SFP+ Cables, XFP Cables, CX4 Cables, Infiniband Cables, etc) and best suppliers for QSFP+ products including 10G & 40G QSFP+ copper and AOC cables. Optics transceivers like QSFP+, SFP, SFP+, XFP, X2, XENPAK, SDH, Bidi are also offered. If you have any interest in our products, you can contact us directly.

Which Cabling Solution is Better for Your Data Center

Although cabling only represents less than 10 percent of the overall data center network investment, it outlives most network elements and treated as the most difficult and potentially costly component. With the datacenter cabling ranging from 1G to 10G, 10G to 40G and even to 100G, more complex cabling is required to ensure a good service or scalability for troubleshooting. In practice, there is no exact solution that will meet all of the cable management needs. However, two kinds of cabling systems can be applied—unstructured system and structured system. Just follow the guidelines and illustration highlighted in the article will go a long way to ensure you with the information required for the successful deployment of a cabling infrastructure in your data center.

Unstructured Cabling System

Unlike the structured cabling system with a managed patch panel, a unstructured cabling only occurs when optical links are deployed point to point or device to device without installing patch panels. In this situation, cabling pathways become congested with an entangled mess of two-fiber optical patch cords. Likewise, routing new patch cords in ceiling or floor trays all the way across a data center each time a new device is deployed is extremely inefficient.

And this entanglement will bring difficulties in routing new patch cords in ceiling or floor trays all the way across the data center whenever a new device is deployed. That greatly influences work efficiency. What’s more, this system causes the overheating of data centers especially around the racks where cable clutter occurs.

Structured Cabling System

Structure cabling emerged as a way to better manage larger data center solution is a big step for the development of optical technology. Structured cabling system is a flexible, reliable and highly efficient for moving, adding and changing the infrastructure as the network grows. This kind of system requires additional investment on pre-terminated MPO cabling such as patch panel to create the cabling infrastructure.

Compared with the unstructured cabling, structured cabling architecture is generally easier to manage and more scalable. And, due to the use of trunked or shared horizontal cabling, it often carries a smaller cable footprint than direct-attach cabling. However, the flexibility of structured cabling presents potential downsides, including cost and link-loss budget. Nevertheless, existing large data centers will likely retain their structured cabling infrastructures, particularly for long-reach, zone-to-zone applications, where it generally remains the more practical choice. The following part will introduce 40G structured cabling solutions.

40G Structured Cabling Solutions

As noted before, structured cabling solutions allow for high consolidation of cabling into a compact patch panel, cabling and connectivity. The traditional duplex multimode SC or LC connections do not support 40G data rate standards, today the MPO technology is commonly found in cassette-based data center installation allowing for easy management and maintenance. Below are cabling solutions of 40G for cable management configurations with the use of MPO patch panel.

One method (seen in the above picture) uses MTP-LC harnesses to transition the MTP connector to LC leads through the use of fiber enclosure loaded with 4 fiber adapter panels (12xMTP Key-up/Key-down). This 12-fiber MTP to LC harness assembly breaks out 4 x LC uniboot legs connecting the SFP+ ports. The lengths of LC harness legs can be customized to adapt to different situations. But this often results in messy cable management. The other method uses MPO/MTP trunk cable and fiber enclosure loaded with 4 MTP high density cassettes (2 x MTP-12 to Duplex LC/UPC 10G OM4) to realize the interconnection. This 96-fiber 1RU rackmount fiber enclosure connects fiber patch cables LC to LC and MTP trunk cable. This method is specially used when the 4xLC ports are not located in close proximity on a single device or are being split between multiple devices. Because it’s more manageable to land the MTP trunk cables into fiber enclosure with individual LC ports for 4xLC patch cables.

Conclusion

Choose the most suitable cabling to support present and future network technology is essential for the long-standing performance of the data center. Structured cabling using an MTP cabling infrastructure is suitable for current 10 Gigabit Ethernet environments while maintaining protection for 40 Gbps environments and beyond. Compared with unstructured cabling, it might be a better solution for you. Except for the right knowledge of a structure cabling, the right tools, patience and discipline are also the key factors that will attribute to the masterpiece of your cable management in data center.

10GBASE-T – Will It be the Best Media Options for 10G Data Center?

Ratified in 2006, 10GBASE-T is the standard to provide 10Gbqs connections over balanced twisted-pair copper, including Category 6A unshielded and shielded cabling. It provides great flexibility in network design due to its 100-meter reach capability. An immediate use for 10GBASE-T is to build the data center access-layer network that connects servers to access switches. But is it the best options for 10G data center? Understanding this requires an examination of the pros and cons of current 10GbE media options.

10GBASE-CX4
10GBASE-CX4 was the first favorite for 10GbE deployments, however its adoption was limited by the bulky and expensive cables, and its reach is limited to 15 meters. The large size of the CX4 connector prohibited higher switch densities required for large scale deployment. Larger diameter cables like 10GBASE-CX4 are purchased in fixed lengths resulting in challenges to manage cable slack. As a result, pathways and spaces may not be sufficient to handle this larger cable.

SFP+ 
SFP+’s support for both fiber optic cables and DAC which makes it a better solution than CX4. SFP+ is commonly used for 10G today, but it has limitations that will prevent itself from moving to every server. The following image shows a SFP+ nodule, SFP+ DAC cable and a 10GBASE-T SFP+ port media converter.

10GBASE-SR—10GBASE-SR (SFP+ fiber) fiber is great for its low latency and longer distance (up to 300 meters), but it is expensive. SFP+ fiber offers low power consumption, but the cost of laying fiber networking everywhere in the data center is prohibitive. The SFP+ fiber electronics can be four to five times more expensive than their copper counterparts, meaning that ongoing active maintenance, typically based on original equipment purchase price, is much more expensive. In addition, replacing a copper connection that is readily available in a server to fiber creates the need to purchase not only the fiber switch port, but also a fiber NIC for the server. EX-SFP-10GE-SR is compatible Juniper SFP+ transceiver that requires a OM3 cable to realize its 10G connectivity, which is an indispensable component for a 10G network.

10GBASE-SFP+ DAC—DAC is a lower cost alternative to fiber, but it can only reach 7 meters and it is not backward-compatible with existing GbE switches. Take MA-CBL-TA-1M as an example, it is designed to cover a distance of 1m for 10G connectivity. The DAC cables are much more expensive than structured copper channels, and cannot be field terminated. This makes DAC more expensive than 10GBASE-T. The adoption rate of DAC will be low since it does not have the flexibility and reach of 10GBASE-T.

10GBASE-T
The major benefit of 10GBASE-T is that it offers the most flexibility, the lowest cost media, and is backward-compatible with existing 1GbE networks. Like all BASE-T implementations, 10GBASE-T covers a lengths up to 100 meters, which gives network designers a far greater level of flexibility in connecting devices in the data center and the most flexibility in server placement since it will work with existing structured cabling systems. For higher grade cabling plants (category 6A and above), 10GBASE-T operates in low power mode on channels under 30 m. This means a further power savings per port over the longer 100m mode. And because 10GBASE-T is backward-compatible with 1000BASE-T, it can be deployed in existing 1GbE switch infrastructures in data centers that are cabled with CAT6 and CAT6A (or above) cabling, enabling network designers to keep costs down while offering an easy migration path to 10GbE.

One challenge with 10GBASE-T is that the early physical layer interface chips (PHYs) consumed too much power for widespread adoption. But there comes a good news with 10GBASE-T is that the PHYs benefit greatly from the latest manufacturing processes. The newer process technologies will reduce both the power and cost of the latest 10GBASE-T PHYs. The latest 10GBASE-T adapters require only 10 W per port. Further improvements will reduce power even more. In 2011, power dropped below 6 W per port, making 10GBASE-T suitable for motherboard integration and high-density switches.

Conclusion
Of all the media options offered above, 10GBASE-T breaks through important cost and power consumption barriers in 10GbE deployment as well as its backwards compatibility with 1GbE networks. Deployment on 10GBASE-T will simplify data center infrastructures, making it easier to manage server connectivity while delivering the bandwidth needed for heavily virtualized servers and I/O-intensive applications. I must say, 10GBASE-T will be the best option for 10GbE data center cabling in the near future.

Fibre Channel Over Ethernet in the Data Center

Ethernet and Fibre Channel are the two typical networks that data center will utilize to present operational and maintenance issues. Traditional Ethernet is a family of frame-based computer networking technologies for local area networks (LANs), whereas Fibre Channel is used for storage area networking (SANs). Fiber over Ethernet (FCoE) is a storage networking protocol that supports Fibre Channel over Ethernet. This article will provide an overview of FCoE, describe the importance to data center that make up the new ecosystem, and explains how the technology is developed today.

What Is FCoE?
Today’s network use different protocols to send information between devices, FCoE is a newly proposed standard that encapsulates Fibre Channel frames into an Ethernet frame at the server (Figure 1), allowing them to run alongside traditional Internet Protocol (IP) traffic. The server encapsulates Fibre Channel frames into Ethernet frames before sending them over the LAN and de-encapsulates them when FCoE frames are received. Server I/O consolidation combines the NIC and host bus adapter (HBA) cards into a single converged network adapter (CNA), which reduces server cabling and power/cooling needs. At present, the Ethernet frame is removed at the Ethernet edge switch to access the Fibre Channel frame, which is then transported to the SAN.

FCoE combined with the advent of 10 Gigabit Ethernet (10 GE) fabrics will grant companies the ability to consolidate their I/O, cables, and adapters while at the same time increase the utilization of their servers. Conceptually FCoE can be broken down into three components: encapsulation of a Native Fibre Channel Frame into an Ethernet frame , the extension of Ethernet to become a lossless fabric, the replacing of a Fibre Channel link with MAC addresses in a lossless Ethernet fabric.

Why Is FCoE Important to the Data Center? 
I/O consolidation is simple in concept: the sharing of both Fibre Channel and Ethernet traffic on the same physical cable or in cases that network isolation is desired, the flexibility to use and configure the same hardware for either type of network load. The benefits end-users will realize from this simple idea are significant. Companies that leverage I/O consolidation will be able to realize significant gains in server slot efficiency with the use of multi-function network/storage. These benefits are further detailed below.

I/O consolidation means a customer can use multi-function network/storage adapters in place of single-function network-specific and storage-specific cards, thereby reducing the number of server slots and switch ports, as well as reducing the number of power consumed for I/O and necessary cooling. This also results in fewer points of management administrators will have to control. A reduction in NICs through I/O consolidation has an additional important advantage. The ability to cool a set amount of heat generated per rack is the primary barrier to data-center expansion and inefficiency encountered today. Reducing the amount of NICs in servers can reduce the amount of heat those servers generate.

Looking Ahead 
As enterprise data centers converge Ethernet and Fibre Channel networks, they can improve performance and reduce power consumption, infrastructure complexity, and cost. The Third generation FCoE architecture with the exception that the core switch now forwards the FCoE frame directly to storage where the Fibre Channel frame is accessed. This architecture solution reduces the server interconnect cabling and adapter card number by at least 50 percent, eliminates the Fibre Channel HBA to SAN optical fiber trunk cable and eliminates the core switch to SAN director fiber trunk cable. Optical connectivity shall be in accordance with IEEE 802.3ae (10GBASE-SR) utilizing OM3 optical fiber.

Summary
FCoE offers a data center unified fabric solution that simplifies operational and maintenance of the cabling infrastructure. FCoE facilitates utilization of low-cost Ethernet electronics and OM3 or OM4 optical connectivity to support 10, 40 and 100 Gigabit data rates. Fiberstore has great advantage of 1000BASE-T/SX/LX SFP, BiDi SFP, 10GBASE-SR/LR SFP+, DWDM SFP+, whole series 40G QSFP+ optics and cables, as well as the 100G CFP2 and CFP4, etc. which help you solve the cost issue in fiber project. Especially the 40G QSFP optics, with the passive optic design, they can be compatible with all the equipment of all major brands.

Reference:
http://www.cisco.com/c/dam/en/us/solutions/collateral/data-center-virtualization/ieee-802-1-data-center-bridging/white_paper_FCIAFCoE.pdf

Design Consideration for 40G Ethernet Network

With the speed in the data center now increases from 10G to 40G, different optical technology and cabling are required. But at first we should figure out the design of 40G Ethernet network. There are several key factors that may affect the transition to 40G. This article today will pay special attention to those aspects that influence data center design consideration.

General Data Center Design
The principal goals in data center design are flexibility and scalability, which involve site location, building selection, floor layout, electrical system design, mechanical design and modularity. Furthermore the key to a successful data center facility: one that is sustainable in the long term; the other is to consider it as a receptacle for equipment and operations, as well as an integrated system, in which each component must be considered to be flexible and scalable. Figure 1 shows a typical data center infrastructure design utilizing preterminated optical solutions.

What Does MPO Connector Means for 40G Data Center?
While speeds have increased to 40G, optical connectivity has remained in a duplex format, whether SC or LC. With the advent of 40G/100G Ethernet, multi-fiber push-on (MPO) connector technology are now used at the electronics interface and further into the data center infrastructure design. MPO technology has displayed proven value in cassette-based data center physical layer installations.

The MPO is defined by TIA-604-5-C, Fiber Optic Connector Intermateability Standard. Type MPO (FOCIS-5) as an array connector that can support up to 72 optical fiber connections in a single connection and ferrule. While the MPO is versatile in the fiber count supported, the 12-fiber MPO is the version widely deployed. Many data center designs today use cassette-based duplex LC connectivity or MPO to duplex LC harnesses at the electronics interface, while 12-fiber MPO-based connectivity is used to connect the trunk cabling to each cassette or harness.

40G Standard Provision
The Habtoor STFA Soil Group (HSSG) has designated 40G to support high-performance computing clusters, blade servers, SANs and network-attached storage. For 40G deployment, the QSFP transceiver will utilize a 12-fiber MPO. Deployment of 40G over multi-mode fiber will be achieved with 4-Tx and 4-Rx fibers from the 12-fiber MPO (see in Figure 2). Each of these four “channels” will transmit 10G for the combined 40G transmission. Single-mode fiber transmission will remain duplex connectivity using course wavelength division multiplexing. The HSSG has also defined the transmission media for 40G to include:

• 40GBASE-SR4 (parallel optics)

100m on OM3/125m on OM4—10G on four fibers per direction

• 40GBASE-LR4(cWDM)

10km on single-mode fiber—4x10G 1300nm wavelength region

• 40GBASE-CR4

7m over copper—4x10G (twinax copper)

• 40GBASE-FR(Serial)

2km on single-mode—4x10G 1550nm

As noted above, the QSFP+ module is specified for use with different standard. The 40GBASE-SR4 is terminated with the MPO connector. For example, Cisco QSFP-40G-SR4 QSFP+ transceiver enables high-bandwidth 40G optical links over 12-fiber parallel fiber terminated with MPO/MTP multifiber female connectors.

For 12-fiber MPO cassette-based optical systems already installed, 40G migration is as simple as removing the existing cassette from the patch panel housings at the equipment and cross connects and replacing the cassette with an MPO adapter panel. Next, an appropriate 12-fiber MPO jumper would be used to cross-connect the trunk cabling as well as interconnect into the QSFP. Though not widely available currently, future preterminated system trunks may utilize 24-fiber MPO connections, both on the trunks and on the cassette. In this case, 40G deployment would require an interconnect harness terminated with two 12-fiber MPO connectors at the QSFP end, and one 24-fiber MPO at the trunk end. This would provide the needed interface with the 24-fiber MPO-based trunk and the 40G QSFP. A 24-fiber MPO jumper would be needed at the system cross connects to ensure polarity was maintained and that skew was within requirements.

Conclusion
The data center infrastructure must be reliable, manageable, flexible and scalable no matter who you are asking for requirements of data center design. It is the responsibility of the network designers to insure best compatibility of data center. As migrating to 40G, we have 40G QSFP and cables within MPO connectivity. Fiberstore supplies a variety of 40G QSFP modules and cables for you to choose from. Besides QSFP-40G-SR4, QSFP-40G-SR4-S and Cisco QSFP-40G-CSR4 are also available. If you are interested in our products, please contact us directly.